summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Merge default into bug23749 (causing no changes)Emile Joubert2013-01-256-44/+309
|\
| * Swap SASL and TLS header codesEmile Joubert2013-01-251-2/+2
| |
| * merge default into bug23749Matthias Radestock2013-01-256-44/+309
| |\
| | * merge default into bug23749Matthias Radestock2013-01-2415-238/+322
| | |\
| | * | s/become/initMatthias Radestock2013-01-241-1/+1
| | | |
| | * | recommend saslMatthias Radestock2013-01-241-1/+1
| | | |
| | * | add some tests for connection refusalMatthias Radestock2013-01-241-4/+21
| | | | | | | | | | | | | | | | | | | | some of this is also tested in the Java client, but I'd rather have it all in one place here.
| | * | improve/fix connection refusalMatthias Radestock2013-01-241-22/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - respond with a 0-9-1 header when the supplied header is not recognised *unless* it is a 1-0 header *and* the 1-0 plug-in is enabled - respond with a 1-0 header when a 1-0 header with an unsupported protocol id is supplied and the 1-0 plug-in is enabled - log a more informative error when a 1-0 header is supplied and the 1-0 plug-in is not enabled
| | * | Reject AMQP 1.0 TLS requests specificallyEmile Joubert2013-01-231-1/+8
| | | |
| | * | Oh yeah, specs.Simon MacMullen2013-01-221-0/+3
| | | |
| | * | Remove knowledge of AMQP frames from the limiter. (attempt 2)Simon MacMullen2013-01-222-9/+25
| | | |
| | * | Merge in defaultSimon MacMullen2013-01-216-49/+294
| | |\ \
| | | * \ merge default into bug23749Matthias Radestock2013-01-186-49/+294
| | | |\ \
| | | | * | refactor: encapsulate rabbit_connection tableMatthias Radestock2013-01-173-5/+12
| | | | | |
| | | | * | cosmeticMatthias Radestock2013-01-171-4/+1
| | | | | |
| | | | * | cosmeticMatthias Radestock2013-01-171-1/+1
| | | | | |
| | | | * | Stick "real" 0-9-1 connections in an ETS table.Simon MacMullen2013-01-173-9/+8
| | | | | |
| | | | * | delay starting of channel_sup_sup as much as possibleMatthias Radestock2013-01-171-13/+13
| | | | | | | | | | | | | | | | | | | | | | | | which makes abandoned connection attempts less costly
| | | | * | merge default into bug23749Matthias Radestock2013-01-178-66/+351
| | | | |\ \
| | | | | * \ Merge headsSimon MacMullen2013-01-171-8/+12
| | | | | |\ \
| | | | | | * | allow 'become' to return, and become something elseMatthias Radestock2013-01-171-8/+12
| | | | | | | |
| | | | | * | | Start the chan_sup_sup later, once we have committed to 0-9-1 - one less hack.Simon MacMullen2013-01-172-22/+19
| | | | | |/ /
| | | | | * | Unwind the stack before taking that one-way trip.Simon MacMullen2013-01-171-2/+4
| | | | | | |
| | | | | * | This guard is not needed.Simon MacMullen2013-01-161-1/+1
| | | | | | |
| | | | | * | UnbreakSimon MacMullen2013-01-161-1/+1
| | | | | | |
| | | | | * | API consistencyMatthias Radestock2013-01-152-3/+3
| | | | | | |
| | | | | * | cosmeticMatthias Radestock2013-01-151-6/+5
| | | | | | |
| | | | | * | neater xref foolingMatthias Radestock2013-01-151-5/+2
| | | | | | |
| | | | | * | merge default into bug23749Matthias Radestock2013-01-152-38/+87
| | | | | |\ \
| | | | | * | | simplify bad_version error handlingMatthias Radestock2013-01-151-6/+5
| | | | | | | |
| | | | | * | | fix typoMatthias Radestock2013-01-151-1/+1
| | | | | | | |
| | | | | * | | Fix the case of sending credit state back when changing credit when there ↵Simon MacMullen2013-01-151-22/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | are no messages in the queue.
| | | | | * | | Separate out can_cons_send.Simon MacMullen2013-01-152-29/+30
| | | | | | | |
| | | | | * | | Specs. And a function I forgot to write.Simon MacMullen2013-01-151-8/+11
| | | | | | | |
| | | | | * | | CosmeticSimon MacMullen2013-01-153-8/+9
| | | | | | | |
| | | | | * | | Allow setting credit for a consumer tag that does not yet exist, since the ↵Simon MacMullen2013-01-142-13/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | adapter wants to do that. We leak if you set credit for a ctag and then don't consume... but the adapter never does that.
| | | | | * | | TODO--Simon MacMullen2013-01-142-5/+8
| | | | | | | |
| | | | | * | | No.Simon MacMullen2013-01-141-1/+0
| | | | | | | |
| | | | | * | | Merge defaultSimon MacMullen2013-01-148-47/+308
| | | | | |\ \ \
| | | | | | * | | Improve workingness further. We now should be handling multiple blocks for ↵Simon MacMullen2013-01-142-67/+76
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | different reasons correctly.
| | | | | | * | | Right. There were a lot of bugs. This fixes most of them. One major ↵Simon MacMullen2013-01-113-46/+55
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | confusion was that both the adapter and the credit impl were trying to normalise against remote count, now it's just the adapter. Also when we get credit we need to unblock. Also there were various things that assumed our local credit could not go negative - well it can and we just need to wait for it to be positive again.
| | | | | | * | | Second attempt at moving credit into the queue. This time we pretend the ↵Simon MacMullen2013-01-104-149/+155
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | limiter is still doing the work. While testing this I note that the credit calculation is crazy when testing with Proton. But it was just as bad before, so let's commit.
| | | | | | * | | consumer_mapping changed since this branch was last worked on.Simon MacMullen2013-01-101-5/+5
| | | | | | | | |
| | | | | | * | | I added that clause early in the history of this bug and it's always been ↵Simon MacMullen2013-01-091-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | wrong. The semantics of that line are just "set is_limit_active to false if the limiter is disabled", is_blocked (which only deals with the channel.flow case anyway) has nothing to do with it.
| | | | | | * | | Stub for 1.0 support.Simon MacMullen2013-01-082-1/+42
| | | | | | | | |
| | | | | | * | | Merge in default.Simon MacMullen2013-01-075-31/+228
| | | | | | |\ \ \
| | | | | | | * | | ResurrectSimon MacMullen2013-01-071-0/+1
| | | | | | | | | |
| | | | | | | * | | And correct specsMichael Bridgen2011-08-241-2/+5
| | | | | | | | | |
| | | | | | | * | | Rectify some poor conflict resolution choices.Michael Bridgen2011-08-242-8/+14
| | | | | | | | | |
| | | | | | | * | | Merge in default. Many conflicts unhappy faceMichael Bridgen2011-08-2331-519/+2196
| | | | | | | |\ \ \
View Lorry Controller Status